Usefulness and Limitations of Polymerase Chain Reaction (PCR) for the Diagnosis and Management of Toxoplasmosis Following Allogeneic Hematopoietic Cell Transplant: Single-center Experience With 31 Patients Over 16 Years

Abstract

Background

Toxoplasmosis is an early post-transplant complication in recipients of allogeneic hematopoietic cell transplant (HCT), typically arising from reactivation of latent infection. Toxoplasma gondii polymerase chain reaction (PCR) has improved detection.

Methods

Single-center, retrospective review of allogeneic HCT recipients who developed toxoplasmosis from August 2008 to November 2024.

Results

We identified 31 cases of toxoplasmosis among 1235 HCT recipients. Ten had infection and 21 had end-organ disease. Fever was the most common clinical manifestation (74.2%). Patients with pulmonary or central nervous system disease often lacked organ-specific symptoms. Toxoplasmosis primarily occurred in patients not on prophylaxis (90.3%), at a median of 28 days post-HCT (interquartile range 20-69 days). Whole blood Toxoplasma PCR diagnosed 80.6% cases and showed a cumulative sensitivity of 93.3%. However, PCR was not always positive at symptom onset, and some asymptomatic patients already had end-organ disease at the time of first PCR positivity. Trimethoprim-sulfamethoxazole (TMP-SMX) was the most used treatment (48.4%). Mortality directly attributable to toxoplasmosis was 12.9%, but all-cause mortality was 61.3%.

Conclusions

Toxoplasmosis is an early post-HCT complication with high morbidity and mortality. Prophylaxis is essential. TMP-SMX is effective, but sometimes it is withheld early post-HCT due to potential myelotoxicity. Given the short window between infection and progression to disease, we recommend twice-weekly monitoring with whole blood PCR while off TMP-SMX and early initiation of TMP-SMX post-HCT for Toxoplasma seropositive patients. Atovaquone may be considered as a bridging prophylaxis until TMP-SMX is started, but its absorption may be compromised early post-HCT and breakthrough cases have been reported.

Toxoplasmosis occurs early post-HCT, often with fever in seropositive patients not on prophylaxis. Toxoplasma PCR aids detection but may initially be falsely negative. Positive whole blood PCR warrants imaging for occult pulmonary and CNS disease. Early prophylaxis initiation is recommended.

Latent toxoplasmosis may reactivate after allogeneic hematopoietic cell transplant (HCT) and cause significant morbidity and mortality [1]. The frequency of toxoplasmosis among HCT recipients varies with the seroprevalence of the population. A pooled analysis of 2404 Toxoplasma-seropositive HCT recipients reported a median prevalence of 7.5% [1]. Toxoplasmosis occurs as an early post-HCT infection, usually within the first 100 days [2–5]. The main risk factors are pre-HCT recipient Toxoplasma immunoglobulin G (IgG) seropositivity, absence of effective antimicrobial prophylaxis, and impaired T-cell function resulting from serotherapy, CD34+ selected grafts, or intensified immunosuppression for the treatment of graft-versus-host disease (GVHD) [6–10]. Toxoplasmosis disease in HCT recipients most commonly includes cerebral, pulmonary, and disseminated forms [3].

Fever is the most common clinical manifestation of toxoplasmosis in HCT recipients. Respiratory and neurologic symptoms may occur but are often nonspecific [1]. Historically, most cases were diagnosed postmortem [3]. Currently Toxoplasma gondii polymerase chain reaction (PCR) testing, which can be performed on whole blood, bronchoalveolar lavage (BAL) fluid, cerebrospinal fluid (CSF), and tissue specimens, is most used. Preemptive serial monitoring of whole blood PCR is recommended in seropositive HCT recipients to identify asymptomatic infection before the development of end-organ disease [4, 8, 9, 11, 12]. However, there are instances of toxoplasmosis disease occurring concurrently with the first positive blood PCR or without any positive results from blood PCR [4, 5, 9, 13, 14].

Here, we present our single-center experience with toxoplasmosis over the 16 years we have been using PCR. Our report provides an updated understanding of patient presentations and disease progression with PCR testing.

METHODS

We conducted a retrospective review of all patients who received allogeneic HCT at our institution from 1 August 2008 (when Toxoplasma PCR became an orderable clinical test) to 30 November 2024. An institutional data query tool (Biomedical Translational Research Information System) was used to determine the number of patients who received allogeneic HCTs within the study timeframe and to determine their pre-HCT Toxoplasma IgG serostatus. Toxoplasma-specific IgG antibody testing was performed at a reference laboratory that used a commercial immunoassay.

Patients with any positive Toxoplasma PCR or pathology suggestive of toxoplasmosis infection or disease were obtained from the Department of Laboratory Medicine database and the Transplant Infectious Diseases group. For those patients with toxoplasmosis, in-depth chart review was performed. All patients were consented for HCT under institutional review board-approved protocols, which permitted the collection of infection data. This analysis of toxoplasmosis was conducted as a quality improvement initiative.

We follow previously described definitions for toxoplasmosis infection and disease [7, 8]. In brief, toxoplasmosis infection required positive blood PCR in the absence of organ involvement, with or without fever. Toxoplasmosis disease was classified as definite, probable, or possible. Definite disease required histo- or cytopathologic confirmation. Probable disease required clinical and radiographic evidence of organ involvement with at least 1 positive PCR from blood, CSF, or the respiratory tract, and exclusion of an alternative diagnosis. Possible toxoplasmosis, where there is no microbiologic evidence, is not included in this manuscript.

T gondii PCR testing was instituted at our center in 2008 and has evolved over the years. The assay is performed on whole blood, CSF, ocular fluid, BAL, induced sputum, and tissue. The initial assay (2008–2018) was performed on 10 µL of nucleic acid extracted on the Roche MagNA Pure automated extraction instrument using the LC Total Nucleic Acid Isolation kit. For the real-time PCR assay, a set of primers and fluorescence resonance energy transfer probes were used targeting a 133-bp fragment of the REP-529 repetitive sequence [15]. PCR was performed on the LightCycler 2.0 instrument (Roche, Inc.). In 2018, the assay was modified to include the B1 gene target in addition to REP-529 due to reports of the parasite genotype affecting the sensitivity of the latter target. The current assay uses 5 µL of nucleic acid extracted on automated extractors (QiaSymphony for whole blood and EasyMag for all other specimen types). Following extraction, a multiplex PCR targeting the REP-529 target (112-bp fragment) and the B1 target (98-bp fragment) in a single tube is performed as a TaqMan real-time PCR reaction. In our study, the B1 target was never detected without concurrent REP-529 target detection. This assay was designed as a qualitative real-time PCR. Although quantitative PCR would be valuable for prognostic assessment, the lack of calibrated standards for this parasite at the time of assay development precluded the provision of quantitative results.

During the early study period, Toxoplasma PCR testing was ordered on an ad hoc basis. Beginning in 2014, whole blood Toxoplasma PCR screening was implemented for most Toxoplasma seropositive patients. Most Toxoplasma seropositive patients who had serial blood PCR evaluation were monitored once to twice weekly while off trimethoprim-sulfamethoxazole (TMP-SMX) prophylaxis. After starting TMP-SMX, monitoring practices varied, though most continued serial PCR monitoring through day +100, during which patients remained local for routine follow-up. Beyond day +100, monitoring was at the provider's discretion, typically reserved for those on intensified immunosuppression or alternative prophylaxis.

At our institution, recommended toxoplasmosis prophylaxis post-HCT is TMP-SMX. TMP-SMX is started postengraftment at the discretion of the transplant physician based on graft function and the patient's tolerance. When TMP-SMX is contraindicated, atovaquone is used as an alternative prophylaxis.

RESULTS

A total of 1235 allogeneic HCTs were performed during the study period. Pre-HCT Toxoplasma IgG serology results were available for all patients, of which 469 (38.0%) were seropositive. Fifty of these patients had both seropositive and seronegative Toxoplasma IgG results pre-HCT. Patients with mixed serology were classified as seropositive, reflecting a clinically conservative approach consistent with our practice, although none of these patients developed toxoplasmosis. Among them, 30 patients initially tested negative but later showed positive serology; 14 transitioned from positive to negative serology; and 6 exhibited fluctuating results, alternating between positive and negative over multiple assessments. Notably, one half of these patients (25 of 50) had low-level positive serology values within 8 IU/mL of the assay's reference threshold, which may reflect analytical uncertainty. Other potential explanations for variable serologic patterns include true seroconversion because of primary infection, false-positive results associated with intravenous immunoglobulin administration, and false-negative results because of immunosuppression.

Thirty-one recipients of allogeneic HCT were diagnosed with toxoplasmosis infection (n = 10) or disease (n = 21), for a prevalence of 2.5%. Seven patients were diagnosed with toxoplasmosis from 2008 through 2013 (prevalence 1.5%, 7/461), during which time Toxoplasma PCR was ordered on an ad hoc basis, whereas 24 patients were diagnosed from 2014 through 2024 (prevalence 3.1%, 24/774) when blood PCR screening was regularly performed.

Baseline Patient Characteristics

Characteristics of patients with toxoplasmosis are summarized in Table 1. Individual patient characteristics are detailed in Supplementary Table 1.

Table 1.

Baseline Characteristics of Patients With Toxoplasmosis

	N (%)
Total allogeneic HCT recipients	1235
Total allogeneic HCT recipients with Toxoplasma IgG seropositivity	469
Total allogeneic HCT recipients with toxoplasmosis	31
Infection	10
Disease	21
Median age at toxoplasmosis diagnosis	40 y (IQR 23-51 y)
Female	17 (54.8%)
Underlying condition
Hematologic malignancy	24 (77.4%)
Severe aplastic anemia	3 (9.7%)
Primary immunodeficiency	2 (6.5%)
Sickle cell disease	1 (3.2%)
Chronic active EBV with HLH	1 (3.2%)
Myeloablative conditioning	9 (29.0%)
Conditioning that included serotherapy	12 (38.7%)
Donor type
Matched related donor	9 (29.0%)
Matched unrelated donor	6 (19.4%)
Mismatched unrelated donor	1 (3.2%)
Haploidentical	12 (38.7%)
Simultaneous haploidentical + cord	2 (6.5%)
Double cord	1 (3.2%)
Stem cell source
Bone marrow	8 (25.8%)
Peripheral blood	20 (64.5%)
Cord blood	1 (3.2%)
Simultaneous cord blood + peripheral blood	2 (6.5%)
CD34+ selection	4 (12.9%)
Pretransplant Toxoplasma IgG (D/R)
D+/R+	11 (35.5%)
D−/R+	17 (54.8%)
D+/R−	1 (3.2%)
D−/R−	1 (3.2%)
D−/R?	1 (3.2%)
PTCy-based GVHD prophylaxis	13 (41.9%)
Receipt of prior allogeneic HCT	5 (16.1%)
Treatment with systemic immunosuppression for GVHD	9 (29.0%)
Relapse of underlying disease	4 (12.9%)

Abbreviations: D, donor; GVHD, graft-versus-host disease; HCT, hematopoietic cell transplant; IgG, immunoglobulin G; IQR, interquartile range; PTCy, posttransplant cyclophosphamide; R, recipient; +, positive Toxoplasma IgG; −, negative Toxoplasma IgG; ?, unknown Toxoplasma IgG.

The median age at toxoplasmosis diagnosis was 40 years (interquartile range [IQR] 23-51 years). A minority of patients were born in the United States of America or Canada (16.1%). Other regions of origin included: Mexico (12.9%), Central America (16.1%), South America (6.5%), Caribbean (25.8%), Europe (6.5%), and Africa (16.1%). Most patients received HCT for hematologic malignancy (77.4%). Myeloablative conditioning was used in 29.0% of cases, 38.7% of patients received serotherapy (antithymocyte globulin or alemtuzumab) as part of conditioning, and 12.9% received CD34+ selected grafts. Donors were HLA-haploidentical in 38.7% and HLA-mismatched unrelated donors in 3.2%. The most frequent stem cell source was peripheral blood (64.5%). Five patients developed toxoplasmosis after a second allogeneic HCT.

Twenty-eight patients (90.3%) were Toxoplasma IgG positive before HCT; therefore, the prevalence of toxoplasmosis among seropositive recipients was 6.0% (28/469). One patient did not have Toxoplasma serology pre-HCT. Two patients (6%) had negative pretransplant Toxoplasma serology. They had received extensive anti-B-cell therapy for their underlying malignancy, which may have contributed to false-negative serology.

Thirteen (41.9%) patients received posttransplant cyclophosphamide (PTCy) as GVHD prophylaxis. At the time of toxoplasmosis, 9 patients (29.0%) were receiving systemic steroids for GVHD, and 4 (12.9%) had relapse of their underlying disease.

Clinical Presentation

Characteristics of toxoplasmosis are summarized in Table 2.

Table 2.

Toxoplasmosis Clinical Presentation

	N (%)
Symptomatic with toxoplasmosis at time of diagnosis	23 (74.2%)
Median time to onset of toxoplasmosis symptoms following HCT	26 d (IQR 16-73 d)
Fever	23 (74.2%)
Manifestations of pulmonary toxoplasmosis	N = 12
Cough	2
Hypoxia	2
Absence of respiratory symptoms	8
Manifestations of CNS toxoplasmosis	N = 13
Altered mental status	3
Headache	1
Absence of CNS-specific symptoms	9
Manifestations of disseminated toxoplasmosis	N = 7
Fever or symptoms specific to affected organ	5
HLH	1
Asymptomatic	1
Pulmonary radiographic findings
Multifocal ground glass opacities and consolidations	8
Multifocal subcentimeter pulmonary nodules	5
CNS radiographic findings
Multiple intracranial lesions	8
Singular intracranial lesion	2
Leptomeningeal enhancement	2
Normal MRI of the brain	2
Receipt of immunosuppressive therapy within 30 d of toxoplasmosis	31 (100%)
Absence of toxoplasmosis antimicrobial prophylaxis	28 (90.3%)
Median absolute lymphocyte count (ALC)a	0.19 K/µL (IQR 0.10-0.47 K/µL)
Median absolute CD4b	87/µL (IQR 76-93/µL)
Median CD4%	19.8% (IQR 15.4%-30.0%)

Abbreviations: CNS, central nervous system; HLH, hemophagocytic lymphohistiocytosis; IQR, interquartile range.

^aALC calculated from the day of toxoplasmosis diagnosis for patients who were postneutrophil engraftment and had white blood cell count > 0.20 ×10⁹/L. Data available for 25 patients.

^bCD4 values calculated for patients with toxoplasmosis diagnosed following day +31 and CD4 values available within 2 weeks of toxoplasmosis diagnosis. Data available for 9 patients.

Ten (32.3%) patients had toxoplasmosis infection, of whom 7 (70%) had fever and 3 were asymptomatic. Twenty-one patients had toxoplasmosis disease—4 (12.9% of the total 31 patients) had definite toxoplasmosis disease and 17 (54.8%) had probable disease. Sixteen of the 21 patients with disease (76%) had fever. Hence, the presence or absence of fever is not helpful to distinguish between infection and disease.

Toxoplasmosis occurred early after transplant. Among symptomatic patients (n = 23, 74.2%), symptom onset occurred a median of 26 days (IQR 16-73 days) following HCT. One-fourth of patients (n = 8, 25.8%) had no symptoms directly attributable to toxoplasmosis at diagnosis by PCR positivity. Of these 8 patients, 3 had infection and 5 had disease. This illustrates the limitations of current PCR assays as a tool for preemptive management of toxoplasmosis. Two patients with infection remained asymptomatic, whereas the third had longstanding multifactorial fevers unchanged by toxoplasmosis diagnosis and treatment. Among the 5 patients with disease who initially had no symptoms attributable to toxoplasmosis, 3 developed fever following toxoplasmosis diagnosis, 1 remained asymptomatic, and 1 had longstanding multifactorial fevers unchanged with toxoplasmosis diagnosis and treatment.

All patients diagnosed with toxoplasmosis had a chest computed tomography (CT) scan to evaluate for pulmonary disease. Among patients with probable pulmonary toxoplasmosis (n = 12), most lacked respiratory symptoms (n = 8, 67%). Radiographic findings were nonspecific and included multifocal ground glass opacities and consolidations (n = 8) and multifocal subcentimeter pulmonary nodules (n = 5) (Figure 1).

Figure 1.

Radiographic spectrum of pulmonary toxoplasmosis. (A-E) Computed tomography of the chest findings showing varied presentations of multifocal ground-glass opacities and consolidations associated with pulmonary toxoplasmosis. (F) Representative computed tomography of the chest showing subcentimeter pulmonary nodule (circled) associated with pulmonary toxoplasmosis.

Almost all patients diagnosed with toxoplasmosis had brain imaging to evaluate for central nervous system (CNS) disease. Twenty-eight (90.3%) had magnetic resonance imaging (MRI) and 1 patient had a CT scan. Of the other 2, 1 was transitioned to comfort care because of malignancy progression and the other could not tolerate the examination. Thirteen patients (41.9% of the total 31) had evidence of CNS toxoplasmosis by imaging. Most of them (n = 9, 69%) did not have any CNS-specific symptoms. Multiple intracranial lesions (n = 8) were more common than single lesions (n = 2). Intraparenchymal brain lesions were often associated with vasogenic edema, and they were found in cerebral hemispheres, midbrain, brainstem, cerebellum, occipital and parietal lobes, and the cervicomedullary junction. Two cases had leptomeningeal enhancement, for which this was the sole radiographic finding in 1 patient. Two cases had unrevealing MRI of the brain but positive CNS Toxoplasma PCRs.

Seven patients had disseminated toxoplasmosis. Most (n = 6) presented with fever and/or organ specific symptoms: 1 with isolated bone marrow involvement and 5 with CNS plus pulmonary disease. The patient with isolated bone marrow involvement presented with hemophagocytic lymphohistiocytosis, as has been previously described [16, 17].

Only 3 (9.7%) patients were taking toxoplasmosis prophylaxis at the time of diagnosis. Two were receiving atovaquone. One patient had concurrent gastrointestinal GVHD (although on tapering oral immunosuppression), which may have impaired atovaquone absorption, whereas the other patient did not have risk factors for impaired absorption. The third patient was prescribed 1 double-strength (DS) TMP-SMX tablet 3 times per week, but there were concerns about medication adherence. Preceding toxoplasmosis, this patient also had relapse of nodular sclerosing Hodgkin lymphoma, for which he was being treated with corticosteroids, brentuximab, and nivolumab.

Most patients were lymphopenic (median 0.19 K/µL, IQR 0.10-0.47 K/µL) and had low CD4 counts (median absolute CD4 was 87/µL, IQR 76-93/µL; median CD4% was 19.8%, IQR 15.4-30.0) at the time of diagnosis.

Diagnostic Methods

Toxoplasmosis diagnostics and classification of infection and disease are summarized in Table 3.

Table 3.

Toxoplasmosis Diagnostics, Classification, Treatment, and Outcomes

	N (%)
Median time to toxoplasmosis diagnosis following HCT	28 d (IQR 20-69 d)
Test establishing diagnosis
Blood PCR	25 (80.6%)
BAL PCR	3 (9.7%)
Brain biopsy PCR	1 (3.2%)
Autopsy	1 (3.2%)
Karius next-generation sequencing	1 (3.2%)
Classification of toxoplasmosis
Infection	10 (32.3%)
Infection symptomatic with fever	7 (22.6%)
Infection with no toxoplasmosis-specific symptoms	3 (9.7%)
Disease	21 (67.7%)
Definite disease	4 (12.9%)
Probable disease	17 (54.8%)
Sites of microbiologically confirmed disease	19 (61.3%)
CNS only	6 (19.4%)
Pulmonary only	7 (22.6%)
Bone marrow only	2 (6.5%)
CNS + pulmonary	4 (12.9%)
Bone marrow + CNS	0 (0%)
Sites of microbiologically confirmed AND radiographically and clinically suspected disease	21 (67.7%)
CNS only	7 (22.6%)
Pulmonary only	7 (22.6%)
Bone marrow only	1 (3.2%)
CNS + pulmonary	5 (16.1%)
Bone marrow + CNS	1 (3.2%)
Cumulative sensitivity of blood Toxoplasma PCR to detect toxoplasmosis	93.3% (28/[28 + 2])
False-negative blood Toxoplasma PCR when symptomatic	5
Infection	1
CNS only	2
Bone marrow only	1
CNS + pulmonary	1
True-positive blood Toxoplasma PCR when asymptomatica	8
Infection	3
Disease	5
Treatment regimens
TMP-SMX	15 (48.4%)
Pyrimethamine-sulfadiazine	5 (16.1%)
Other combination therapiesb	10 (32.3%)
No treatment (postmortem diagnosis)	1 (3.2%)
Median treatment duration for infection	30 d (IQR 17 to 40 d)
Median treatment duration for disease	49 d (IQR 42 to 127 d)
Mortality attributable to toxoplasmosis	4 (12.9%)
All-cause mortality	19 (61.3%)

Abbreviations: BAL, bronchoalveolar lavage; CNS, central nervous system; IQR, interquartile range; PCR, polymerase chain reaction; TMP-SMX, trimethoprim-sulfamethoxazole.

^aAsymptomatic or did not have a change in clinical status attributable to toxoplasmosis.

^bCombination therapies included the following: TMP-SMX + pyrimethamine; TMP-SMX + sulfadiazine; TMP-SMX + pyrimethamine + clindamycin; pyrimethamine + clindamycin; pyrimethamine + atovaquone; atovaquone + azithromycin + clindamycin; atovaquone + azithromycin; atovaquone + clindamycin; atovaquone + sulfadiazine; azithromycin + clindamycin.

Toxoplasmosis diagnosis occurred a median of 28 days (IQR 20-69 days), range 8 to 405 days, following HCT (Figure 2). Fifteen cases occurred on or before day +28 post-HCT.

Figure 2.

Days from hematopoietic cell transplant to toxoplasmosis diagnosis.

The diagnosis was established by whole blood Toxoplasma PCR in 25 (80.6%) patients. Other diagnostic sources for PCR included BAL (n = 3, 9.7%) and brain (n = 1). One patient was diagnosed at autopsy, and 1 by next-generation sequencing ([NGS], Karius).

Ultimately, 28 patients developed positive whole blood Toxoplasma PCRs. Twenty-two patients with positive blood Toxoplasma PCRs had a negative PCR preceding the first positive result, a median of 7.0 days (IQR 4.3-7.0 days) earlier. There was a median of 2 days (IQR 1.8-5.3 days) from the time of symptom onset to the first positive blood Toxoplasma PCR. Once positive, the median duration of blood PCR positivity was 8 days (IQR 7-18 days). Supplementary Figure 1 shows the cycle threshold values of positive blood PCRs for each patient.

The PCR was positive in respiratory samples from 11 patients, leading to the diagnosis of probable pulmonary toxoplasmosis: 8 BAL specimens, 2 induced sputum specimens, 1 pleural fluid. One additional patient with positive blood Toxoplasma PCRs had suspected pulmonary toxoplasmosis based on the presence of pulmonary nodules that responded to toxoplasmosis treatment, but BAL was not performed.

Toxoplasma PCR was positive in the CSF of 9 patients. One additional patient had a specialized research assay that was positive from the CSF. Three additional patients with other sites of toxoplasmosis disease or blood PCR positivity had suspected CNS toxoplasmosis based on radiographic identification of brain parenchymal lesions that responded to toxoplasmosis treatment. Among the 13 patients with CNS toxoplasmosis, the median time from HCT to diagnosis was 92 days (IQR 51-119 days). In 6 patients (IDs 2, 4, 5, 7, 12, 13), the CNS was the sole site of disease; 2 of these 6 patients had negative blood PCRs (1 of whom had serial evaluation and 1 of whom had a single evaluation). Another patient (ID 10) with CNS and pulmonary disease had negative blood PCRs on serial evaluation. The remaining 10 CNS cases ultimately had positive blood PCRs. Further CNS diagnostic details are presented in Supplementary Table 2.

Performance of Whole Blood Toxoplasma PCR

Excluding 1 patient with a postmortem diagnosis of toxoplasmosis who did not have serial whole blood Toxoplasma PCR testing, blood Toxoplasma PCRs remained negative in only 2 patients with toxoplasmosis—sites of disease were CNS (n = 1) and pulmonary plus CNS (n = 1). In all other patients, whole blood PCR ultimately became positive. Thereby, the cumulative sensitivity of the whole blood Toxoplasma PCR to detect toxoplasmosis at any time point during the course was 93.3%.

However, the sensitivity of blood PCR was lower at the onset of symptoms. Five patients had negative blood Toxoplasma PCRs while symptomatic with fever, which was ultimately attributable to toxoplasmosis. As stated previously, 2 of these patients with disease—1 with CNS and the other with both pulmonary and CNS involvement—never developed positive blood PCRs. The other 3 patients had a range of 1 to 3 days of symptoms preceding 1 negative blood PCR. One of these patients had just infection with fever, 1 had CNS disease, and 1 had bone marrow involvement. For these 3 patients, the subsequent blood Toxoplasma PCRs were all positive, at a range of 6 to 8 days following the onset of symptoms attributable to toxoplasmosis.

Whole blood Toxoplasma PCR did not reliably detect infection before disease occurred. Blood Toxoplasma PCR was positive for 8 patients who were asymptomatic or did not have a change in clinical status attributable to toxoplasmosis, but 5 of them already had end-organ disease, whereas 3 patients had infection only. Sites of organ disease included pulmonary (n = 2), CNS (n = 1), pulmonary plus CNS (n = 1), and bone marrow plus CNS (n = 1). This limits the usefulness of PCR for the “preemptive” management of toxoplasmosis.

Treatment and Outcomes

Toxoplasmosis treatment and outcomes are summarized in Table 3.

One patient who never received treatment died from toxoplasmosis that was diagnosed postmortem. Treatment duration for the other 30 patients varied. Fourteen patients died while on therapy for toxoplasmosis, although death was attributed to toxoplasmosis in only 3 of them (median treatment duration for these 3 patients was 43 days, range 39-154 days). Sixteen patients completed a course of treatment: 7 were treated for infection and 9 were treated for disease. The median treatment duration for infection was 30 days (IQR 17-40 days). The median treatment duration for disease was 49 days (IQR 42-127 days).

Fourteen patients (45.2%) were treated exclusively with TMP-SMX. Another patient was treated with a TMP-SMX-based regimen for more than half the treatment duration. TMP treatment doses ranged from 10 to 15 mg/kg/day. Only 1 patient with infection was treated with low-dose TMP-SMX (1 DS tablet daily). None of these patients had antimicrobial treatment failure. Five patients (16.1%) were treated with pyrimethamine and sulfadiazine-based regimens. The other 10 patients were treated with combination regimens that included the following: TMP-SMX, pyrimethamine, sulfadiazine, atovaquone, clindamycin, and azithromycin.

All patients who completed a treatment course, except 1, were transitioned to secondary prophylaxis following treatment. The most frequent secondary prophylaxis regimen was TMP-SMX DS 3 times weekly (n = 10), followed by TMP-SMX DS once daily (n = 4) and atovaquone 1500 mg/day (n = 1).

Continued PCR monitoring during treatment helped assess response, detect relapse, and facilitate TMP-SMX dose titration to minimize toxicity. Two patients on TMP-SMX (patients 25 and 28) had Toxoplasma PCR positivity relapse within 1 month of initial PCR positivity. Patient 25 relapsed 21 days after clearance because of TMP-SMX nonadherence. Patient 28 had isolated recrudescent PCR positivity 28 days after clearance on TMP-SMX 10 mg/kg/day. Before the positive PCR resulted, TMP-SMX had been decreased to 1 DS tablet once daily, and the PCR reverted to negative within 5 days on the lower dose.

Two other patients (patients 13 and 24) experienced delayed recrudescence of Toxoplasma PCR positivity. Patient 13 had PCR positivity 93 days after the initial positive result in the setting of chronic GVHD and nonadherence with TMP-SMX prophylaxis. This same patient experienced another recrudescence of toxoplasmosis infection 668 days after clearing the original blood PCR, again in the setting of ongoing chronic GVHD. Patient 24 experienced recrudescent toxoplasmosis 153 days after initial clearance while off effective prophylaxis and undergoing treatment for liver GVHD.

Three patients died with persistent Toxoplasma PCR positivity, all of whom were started on a second-line regimen of azithromycin and clindamycin, with 2 also receiving atovaquone. One of them was switched to TMP-SMX because of increasing Toxoplasma DNA values (decreasing cycle threshold). There was apparent response based on cycle threshold values, but he developed acute respiratory distress syndrome in the setting of concomitant rhinovirus/enterovirus, adenovirus, probable aspergillosis, and engraftment syndrome, in addition to disseminated toxoplasmosis.

Death directly attributable to toxoplasmosis occurred in 4 patients (12.9%) (the 3 previoiusly mentioned and the 1 diagnosed at autopsy), but all-cause mortality in the cohort was high (n = 19, 61.3%).

DISCUSSION

We present the largest United States single-center cohort of post-HCT toxoplasmosis. Our data provide context for recent European Conference on Infections in Leukemia guidelines for the management of toxoplasmosis [8]. In our cohort, toxoplasmosis emerged very early after HCT, often in hospitalized, newly engrafted patients, with fever as the most common symptom. Pulmonary and CNS disease were common, often without organ-specific symptoms. Whole blood Toxoplasma PCR diagnosed most cases, but we found it to be an imperfect tool for preemptive management, as it could not reliably differentiate between patients with infection and patients with end-organ disease. Additionally, positive results frequently coincided with fever rather than preceding it significantly. Most patients were successfully treated with TMP-SMX. Few deaths were directly attributable to toxoplasmosis, but all-cause mortality was high.

As is well known, toxoplasmosis after HCT represents reactivation of past disease, and there are no well-documented cases of donor-derived toxoplasmosis in HCT [18]. Therefore, most patients at risk may be identified by positive serology. However, 2 of our patients with toxoplasmosis were seronegative, consistent with other reports [11, 14, 19]. Prior B-cell targeting immunosuppression may cause falsely negative serology, so clinicians should consider toxoplasmosis in compatible cases regardless of serology results.

In our cohort, most toxoplasmosis occurred very early post-HCT in patients who were not yet receiving TMP-SMX prophylaxis because of concerns about myelotoxicity (Figure 2). We advocate for initiation of TMP-SMX as soon as possible. If this is not feasible, we recommend using atovaquone starting on day +14, being mindful that atovaquone's absorption may be limited unless taken with fat.

Increased use of serotherapy in conditioning and PTCy-based GVHD prophylaxis may contribute to this early occurrence of toxoplasmosis. Many patients in our cohort received PTCy-based GVHD prophylaxis, and other reports have documented toxoplasmosis in HCT recipients who received PTCy [16, 20, 21]. PTCy has been associated with higher rates of bacterial, viral, and fungal infections because of delayed immune reconstitution [22–28], so further comparative studies are needed to evaluate PTCy as a potential risk factor for toxoplasmosis.

Monitoring with whole blood Toxoplasma PCR may facilitate early diagnosis, but preemptive management is limited. False negatives are possible, and there is minimal lag between PCR positivity and disease onset, as was seen in our cohort and other published studies [4, 5, 9, 13, 14]. Given this short window, we recommend twice-weekly PCR testing for at-risk patients while not on prophylaxis, plus targeted and repeated testing in febrile patients. We continue monitoring even after prophylaxis is initiated, but this may not be cost-effective for patients taking TMP-SMX. Because end-organ disease may already be present at the time of the first positive PCR, an active search for pulmonary and CNS involvement with chest CT and brain MRI is appropriate. Despite its limitations, Toxoplasma PCR-based diagnostics may reduce toxoplasmosis-related mortality [2, 11]. The importance of Toxoplasma PCR monitoring is underscored in our cohort by differences in outcomes in patients with CNS involvement. All 4 patients who died from CNS toxoplasmosis lacked routine blood surveillance (although 1 never had positive blood PCRs). In contrast, among 9 surviving CNS toxoplasmosis patients, 7 had serial monitoring with positive results.

While the inclusion of Toxoplasma PCR has significantly increased the sensitivity of the laboratory diagnosis, the assay performance varies. No commercial assays exist, and most rely on in-house assays targeting REP-529 (200–300-fold repeat) or the B1 gene (35-fold repeat) [29]. Few studies compared PCR performance, but 1 found a 97.8% sensitivity, with repeat testing needed for low parasite levels [29]. Our assay evolved from targeting REP-529 alone to including B1 to address genetic variants, specifically the non-genotype II parasites where the efficacy of REP- 529 target may be reduced [30]. In this study, cumulative whole blood PCR sensitivity was 93.3%, with potential false negatives because of low parasite burden or sequestered disease. The calculated limit of detection of our assay is approximately 100 parasites/mL, which aligns with published studies [31].

Our series includes a case where Karius NGS detected toxoplasmosis before whole blood PCR (with 2 targets) was positive. This patient was symptomatic with fever for 3 days with a negative blood PCR. Two days later, NGS identified T gondii. Three days following NGS, a repeat blood PCR resulted positive. Other studies have also reported NGS use for diagnosing toxoplasmosis [17, 32, 33]. This instance suggests NGS may play a role in toxoplasmosis diagnostics, although widespread availability, cost, and turnaround time remain a barrier.

Treatment of toxoplasmosis early post-HCT with pyrimethamine/sulfadiazine or TMP-SMX is challenging because of the myelotoxicity associated with these regimens. Some experts favor pyrimethamine/sulfadiazine [8], but meta-analyses suggest both are acceptable [34]. We have not had any clinical failures when using treatment dose TMP-SMX, but we continue monitoring during treatment and while patients remain immunocompromised following treatment.

In summary, toxoplasmosis is a relevant and significant cause of morbidity in allogeneic HCT recipients. Toxoplasmosis is most common in the early post-HCT period, often manifest with fever as the sole symptom. Molecular testing remains the cornerstone of diagnostics, although whole blood Toxoplasma PCR can be falsely negative and fail to facilitate preemptive treatment of infection before the development of disease. Despite the limitations of blood PCR, its use seems to be associated with improved treatment outcomes for toxoplasmosis.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.